Lesson 7 of 8
The paper and the portfolio
Where the published returns lived: small companies, thin trading, no analyst coverage, equal weights, hundreds of names. Little of that reaches a person.
By Javier Sanz
Founder of ValueScreener. Built Ninety Nine, a securities services provider, and sold it to a tier-one global bank; ran operations and internationalization at Alpaca, the brokerage API.
Who writes thisEditorial processLinkedIn (opens in a new tab)X (opens in a new tab)
In one minute
- Published results are built from hundreds of positions, rebalanced on a schedule, often equal weighted and often concentrated in small and thinly traded companies.
- Piotroski's own paper reports that the benefit was concentrated in small and medium companies, low share turnover and firms with no analyst following.
- A person holding fifteen names, paying costs and taxes, is running a different object from the one the paper measured.
Every lesson so far has been about whether the result is real. This one assumes it is, and asks a narrower question: how much of it can reach somebody who actually holds shares.
What a published portfolio is, against what a person holds
The paper's portfolio
- Hundreds of positions, formed by rule on a stated date.
- Often equal weighted, so the smallest company carries the same money as the largest.
- Rebalanced annually or monthly, with no commissions, no spread and no tax.
A person's portfolio
- Ten to twenty names chosen one at a time, each researched.
- Weighted by conviction and by what the position costs if it is wrong.
- Every trade pays a spread, and in a taxable account every sale pays tax.
Both can be described as value investing. They are not the same object.
Where the published return actually lived
Piotroski's 2000 paper is the clearest case because it says so itself. Working inside the highest book-to-market quintile between 1976 and 1996, he reports that a strategy buying expected winners and shorting expected losers generated a 23 percent annual return, and that within the high book-to-market portfolio the benefits of financial statement analysis are concentrated in small and medium-sized companies, in companies with low share turnover, and in firms with no analyst following.
That is a precise description of where the money was, and it is also a precise description of what a large fund cannot buy and what a small investor can only buy slowly.
McLean and Pontiff find the same shape across their 97 predictors: returns are higher for portfolios concentrated in stocks with high idiosyncratic risk and low liquidity. The published edge lives in the corner of the market that is hardest to trade.
Four subtractions, in order
From the printed number to what is left
-
Weighting. If the result is equal weighted, a portfolio that respects liquidity will not reproduce it. The paper's own construction tells you how much of the return sat in the smallest names.
-
The short side. A long-short result is two portfolios. Many readers can only hold one of them, and the short half is often where the extra return was.
-
Costs. Every rebalance pays a spread on both sides. A rule with high turnover in illiquid names pays it many times.
-
Tax. In a taxable account, a rule that sells its winners each year converts a paper return into an after-tax return that is lower by whatever the local rate is.
The arithmetic of turnover, once
What survives the trip
Not nothing. The subtractions hit a mechanical, high-turnover reproduction of a published rule hardest, and hit a small number of researched holdings held for years much less. That is the honest argument for the way this site is built: fewer decisions, each one written down, each one checked against filings, with turnover low enough that the costs above stay small.
It is not an argument that the returns in the papers are available. It is an argument about which parts of them survive contact with a real account.
Check yourself
Where did Piotroski report the benefits of his analysis were concentrated?
- In the largest and most liquid companies
- In small and medium-sized companies, low share turnover, and firms with no analyst following
- In companies paying the highest dividends
Show the answer
Answer: B.
His paper states it directly, inside the highest book-to-market quintile between 1976 and 1996.Why does the weighting rule in a study matter to somebody reproducing it?
- It changes the tax rate
- Equal weighting puts as much money in tiny companies as in large ones, which a real portfolio usually cannot do
- It changes the accounting standard used
Show the answer
Answer: B.
Much of a published return can sit in the smallest names, which is exactly where an equal weighted portfolio puts an outsized share of the money.At a round trip cost of 0.6 percent, what does turning a portfolio over twice a year cost?
- 0.3 percent
- 0.6 percent
- 1.2 percent
Show the answer
Answer: C.
Two round trips at 0.6 percent each is 1.2 percent a year, before tax, against an edge that the decay evidence has already reduced.Sources
- Piotroski, Journal of Accounting Research, 2000: the 1976 to 1996 sample inside the highest book-to-market quintile, the 23 percent annual return on the long-short strategy, and the concentration in small and medium firms, low share turnover and no analyst following.
- McLean and Pontiff, "Does Academic Research Destroy Stock Return Predictability?", Journal of Finance, 2016: returns higher for portfolios concentrated in stocks with high idiosyncratic risk and low liquidity.
- Tweedy, Browne, What Has Worked in Investing, 1992, revised 2009, for the equal weighted construction of the tabulated studies.
What this leaves out
- The 0.6 percent round trip and the three turnover levels are illustrative. Real costs depend on the company, the size of the order and the day.
- Tax treatment differs by country and by account type, and nothing here describes any reader's tax position.
- Piotroski's 23 percent is his reported figure for a long-short strategy in his sample period, not a return anyone received.
- Low turnover reduces cost drag. It does not make a method work.
- Nothing here is advice about any security or strategy.